2078
Chem. Mater. 1994,6, 2078-2084
Electrochemical Intercalation of Oxygen in LasNiO4+,: Phase Separation below Ambient Temperature I. Yazdi, S. Bhavaraju, J. F. DiCarlo, D. P. Scarfe, and A. J. Jacobson* Department of Chemistry and Texas Center for Superconductivity, University of Houston, Houston, Texas 77204- 564 1 Received May 10, 1994. Revised Manuscript Received August 5, 1994@’ Electrochemical intercalation of oxygen in LazNiO4+, has been investigated by galvanostatic and potential step experiments for 268 K I T 5 295 K in the composition range 0 Ix I 0.14. At room temperature a single-phase region in the composition range (0.06 < x < 0.14) is observed. When the temperature is lowered t o 273 K, the electrochemical data show evidence for phase separation into three distinct two-phase regions with compositions centered a t x = 0.076, 0.100, and 0.119. Single-phase regions with compositions centered a t x = 0.065,0.087, and 0.112 are observed between the two-phase regions. The kinetics of the phase separation are slow. The results are consistent with recent low-temperature X-ray and neutron diffraction studies of phase separation at specific compositions.
Introduction Intercalation of oxygen in oxides with the perovskite and LazM04 (M = Cu, Ni) structures has recently been shown to occur at ambient temperature. Both chemical and electrochemical oxidation techniques in aqueous KOH electrolytes have been described.l-19 The majority of the studies have examined the LazCu04+, system because of the observation of superconductivity on Abstract published in Advance ACS Abstracts, September 15,1994. (1) Wattiaux, A.; Park, J.-C.; Grenier, J.-C.; Pouchard, M. C. R . Acad. Sci. Paris, Ser. 11, 1990,310,1047. (2) Suryanarayanan, R.; Gorochov, 0.;Rao, M. S. R.; Ouhammou, L.; Paulus, W.; Heger, G. Physica C 1991,185-189, 573. (3)Grenier, J.-C.; Wattiaux, A,; Lagueyte, N.; Park, J.-C.; Marquestaut, E.; Etourneau, J.; Pouchard, M. Physica C 1991,173,139. (4)Wattiaux, A.; Fournes, L. ; Demourges, A.; Bernaben, N.; Grenier, J.-C.; Pouchard, M. Solid State Commun. 1991,77,489. (5)Bezdicka, P.; Wattiaux, A.; Grenier, J.-C.; Pouchard, M.; Hagenmuller, P. Z. Anorg. Allg. Chem. 1993,619,7. (6)Rudolf, P.; Schollhorn, R. J. Chem. SOC.,Chem. Commun. 1992, 1158. (7) Rudolf, P.; Paulus, W.; Schollhorn, R. Adv. Mater. 1991,3,438. (8)Grenier, J.-C.; Lagueyte, N.; Wattiaux, A,; Doumerc, J.-P.; Dordor, P.; Etourneau, J.; Pouchard, M.; Goodenough, J . B.; Zhou, J. S. Physica C 1992,202,209. (9)Grenier, J.-C.; Wattiaux, A.; Doumerc, J.-P.; Dordor, P.; Fournes, L.; Chaminade, J.-P.; Pouchard, M. J.Solid State Chem. 1992,96,20. (10)Chou, F. C.; Cho, J. H.; Johnston, D. C. Physica C 1992,197, 303. (11)Grenier, J.-C.; Wattiaux, A.; Demourges, A.; Pouchard, M.; Hagenmuller, P. Solid State Ionics 1993,63-65, 825. (12)Radaelli, P. G.; Jorgensen, J. D.; Schultz, A. J.; Hunter, B. A.; Wagner, J. L.; Chou, F. C.; Johnston, D. C. Phys. Rev. B 1993,48, 499. (13)Takavama-Muromachi, E.: Sasaki, T.: Matsui. Y. Phvsica C 1993,207,9?. (14)Demourges, A.; Wattiaux, A.; Grenier, J.-C.; Pouchard, M.; Soubevroux. J. L.: Dance. J. M.: Haeenmuller. P. J.Solid State Chem. 1993,305,458. (15) Demourges, A.; Weill, F.; Grenier, J.-C.; Wattiaux, A,; Pouchard, M. Physica C 1992,192,425. (16)DiCarlo, J. F.; Yazdi, I.; Bhavaraju, S.; Jacobson, A. J . Chem. Muter. 1993,5, 1692. (17)Johnston, D. C.; Borsa, F.; Canfield, P. C.; Cho, J. H.; Chou, F. C.; Miller, L. L.; Torgeson, D. R.; Vaknin, D.; Zarestky, J.; Ziolo, J.; Jorgensen, J. D.; Radaelli, P. G.; Schultz, A. J.; Wagner, J . L.; Cheong, S.-W.; Bayless, W. R.; Schirber, J . E.; Fisk, Z. Proceedings of a Workshop on “Phase Separation i n Cuprate Superconductors,”: Coltbus, Germany 4-10 Sept. 1993 Springer-Verlag, to be published. (18)Schwartz, M.;Cahen, D.; Rappaport, M.; Hodes, G. Solid State Ionics 1989,32/33,1137. (19)Scolnik, J.;Rappaport, M.; Hass, N.; Dai, U.; Cahen, D. Physica C 1993,209,1993. @
.
I
’
0897-475619412806-2078$04.50/0
oxidation.’ Complex phase separation occurs as a function of composition and temperature, and two distinct superconducting compositions have been reported.8J0 The phase separation and properties of the LagCu04 system have recently been reviewed-l’ Similar experiments have been carried out on the structurally analogous LazNiO4+, The phase diagram of LazNiOs+, has also been extensively investigated by nonelectrochemical methods and like the copper system shows phase ~ e p a r a t i o n . ~ O - ~ ~ Demourges et al. reported the synthesis of LazNi04+, phases in the composition range 0 5 x I 0.25 by controlled potential electrolysis (600 mV vs Hg/Hg0).14 Specific compositions at x = 0.0, 0.03, 0.11, 0.14, 0.17, 0.18,0.23,and 0.25 were synthesized and characterized by X-ray diffraction and by measurements of their opencircuit potentials. A neutron powder diffraction study of LazNi04.25prepared electr~chemically~~ confirmed the location of the interstitial oxygen atoms in the Lag02 layers, in agreement with previous studies of oxidized samples prepared at high temperature.20 Electron microscopy of La2Ni04,gs revealed the presence of a number of different commensurate and incommensurate superlattices associated with the interstitial oxygen distribution and ~ 0 n t e n t . l ~ DiCarlo et a1.16 examined electrochemical intercalation in LazNi04+, at lower compositions, 0 5 x I0.145, by galvanostatic and potential step methods. At ambient temperature the intercalation-deintercalation reactions were shown to be reversible, and the electrochemical data give direct information about the phase diagram. A single-phase region in the composition range 0.06 < x < 0.145, a two-phase region for 0.01 < x 5 0.06, and a single-phase region for 0 Ix < 0.01 were observed. The results are in excellent agreement with those of Rice and Buttrey obtained by equilibration of LaaNiOw, (20)Jorgensen, J . D.; Dabrowski, B.; Pei, S.; Richards, D. R.; Hinks, D. G. Phys. Rev. B 1989,40,2187. (21)Rice, D. E.;Buttrey, D. J. J.Solid State Chem. 1993,105,197. (22)Tamura, H.; Hayashi, A.; Ueda, Y. Physica C 1993,216,83. (23)Buttrey, D. J.; Honig, J . M. In The Chemistry of High Temperature Superconductors, Rao, C. N. R.. Ed.: World Scientific Publishing: Singapore, 1991;pp 283-305.
0 1994 American Chemical Society
Chem. Mater., Vol. 6, No. 11, 1994 2079
Intercalation of Oxygen in La&iOa+x
crystals in oxygen a t higher temperatures.21 The only discrepancy between the phase diagrams determined by the two methods is that a phase with P42/ncm symmetry in the narrow composition range (0.02 < x < 0.03) was not observed in electrochemical experiments using polycrystalline electrodes. Recent synchrotron X-ray and neutron diffraction measurements indicate that phase separation occurs in the single-phase composition region (0.06 < x 0.145) on cooling to just below room t e m p e r a t ~ r e . ~The ~,~~ initial suggested that LazNi04.105phase separates below room temperature into an orthorhombic phase with a doubled c axis and a tetragonal phase. A 40 50 m 10 more detailed investigation of this and other composiTwo theta tions has subsequently provided a more detailed description of the phase diagram.25 The neutron diffracb. tion data were interpreted in terms of staged layers of intercalated oxygen atoms between LazNi04 layers with the orthorhombic Bmab type of distortion. The oxygen atoms occupy intestitial sites in the La202 layers. The phase at x = 0.105, for example, has been identified as a single-phase stage-2 material. In general, phase separations were observed to occur quite slowly and to show hysteresis. The neutron diffraction s t ~ d i e suggest s ~ ~ ~that ~ ~the phase separation should be apparent in electrochemical measurements a t temperatures below ambient. In this paper, an electrochemical study of the phase diagram 0 of LazNiO4+, a t temperatures between 268 and 295 K Two theta is described and the results compared with the neutron Figure 1. X-ray diffraction patterns of the LazNiO4 starting diffraction measurements.
t
Experimental Section LazNiO4+, was prepared by standard solid state techniques. Stoichiometric amounts of La203 (predried at 800 "C,Aldrich 99.9%) and NiO (Aldrich 99.99%) were ground and heated in air at temperatures of 1250 (24 h), 1250 (24 h), 1270 (48 h), and 1300 "C (12 h). The sample was pelletized and heated to 1380 "C for 12 h. Powder X-ray diffraction data were measured with a Scintag XDS2000 diffractometer using Cu K a radiation in the range 10" 5 2 0 5 70". Measurements were made on the starting material and on electrodes. The electrodes were mounted directly on the diffractometer and diffraction patterns recorded from the front surface. Lattice constants were obtained by fitting the data using GSASZ6"he starting electrode material was tetragonal (IUmmm) with lattice constants a = 3.866(1) and c = 12.665(2)A. At the end of the entire sequence of experiments (-3 months), an X-ray diffraction pattern of the electrode showed no lines due to impurity phases and little evidence for line broadening. Tetragonal lattice constants of a = 3.858(1) A and c = 12.609(1)A were obtained for a partially oxidized sample. The X-ray diffraction patterns for the starting material and final product are shown in Figure 1. Thermogravimetric measurements were performed by reduction of the sample t o nickel metal and lanthanum oxide in a 5%hydrogen-95% nitrogen atmosphere to determine the oxygen content. Samples prepared by the above procedure were used in the electrochemical experiments in the following way. A pellet, 75.70 mg and 8 mm in diameter, was painted on one surface with gold paste, and a 0.025 cm diameter platinum wire was attached to the paint. The gold film was then dried at room temperature for 3 h, annealed in air at 900 "C for 12 h, and (24) Tranauada. J. M.; Buttrey, D. J.; Rice, D. E. Phrs. Rev. Lett. 1993, 70, 44g. (25) Tranauada. J. M.: Kong, Y.: Lorenzo, J. E.: Buttrev, D. J.;Rice,
D. E.; Sacha;, V. Phys. Rev. g, in press. (26)von Dreele, R. B.; Larson, A. C. GSAS Users Guide; Los Alamos National Laboratory: Los Alamos, NM, 1985-1992GSAS.
I1
material and of the LazNiO4 electrode at the end of the lowtemperature experiments.
furnace cooled. Leads attached by this method were shown to behave ohmically by placing two such gold films on either side of a pellet and checking the I-V characteristics. A linear relationship between current and voltage indicated proper ohmic contact. Pellets without gold paint but heat treated in the same way were analyzed thermogravimetrically and had an average composition LazNiOd,ls+o.oz. The microstructures of typical electrodes were examined by scanning electron microscopy using an IS1 SS40 instrument. A typical microstructure is shown in Figure 2 for the surface and for a cross section of an electrode. The LazNiO4 particle sizes are 5-10 rum. Electrochemical experiments were performed using a Biologic MacPile potentiostat-galvanostat and a three-electrode system with the sample as the working electrode, gold foil as the counter electrode, and Hg/HgO/l M KOH (Eo= +0.098 V vs SHE) as the reference electrode. The electrolyte was 1M KOH that had been purged with nitrogen for 3 h prior t o the start of an experiment. The experiments were carried out with a continuous nitrogen purge to avoid the competing reaction of oxygen reduction that can occur at low potentials. The temperature of the cell was lowered in a bath of ethylene glycol and water, cooled by a cold-finger Neslab bath cooler. The reference electrode was at ambient temperature during the experiment. The temperature was maintained at the desired level (k0.2 "C)with a Therm-0-Watch L6.1OOOSS temperature controller and a mercury thermometer.
Results Electrochemical reduction and reoxidation of LazNi04+, were studied by both galvanostatic (constant current) and potential step techniques. Galvanostatic experiments provide an overview of the changes in the composition and kinetics occurring during the intercalation reaction. Potential step experiment^^^ provide
Yazdi et al.
2080 Chem. Mater., Vol. 6, No. 11, 1994
b
.o.d
0 0.02
0.04
0.06 0.08
0.1 0.12 0. 4
x in La2Ni0,+x
Figure 3. Galvanostatic reduction and oxidation curves (50 at 273 K.
4) for LafliO,
Figure 2. Scanning electron micrographs of ( 8 )the top surface and Ihl a cross section of a LanNiO., electrode.
more details of the intercalation chemistry. In potential step experiments, the potential is changed abruptly by a small amount from the equilibrium value, and the current is monitored. The voltage step away from equilibrium initially produces a large current which then decays to a small preset value. When the current falls to the preset limiting current value, the voltage is stepped again. With small voltage steps, the potential step technique provides high composition resolution and data approaching, but not at, true equilibrium. The approach to equilibrium is determined by the value of the limiting current. Data are acquired rapidly because on each step the initial current is high. A galvanostatic experiment camed out a t the limiting current would take much longer. True open-circuit (equilibrium) measurements are even slower and consequently cannot be used to obtain data for many different compositions. The potential step approach is a compromise technique that provides “quasi-equilibrium”data with high composition resolution. Equilibrium data can be used to determine the number of phases present in specific composition ranges for the pseudobinary system LazNi04h. The approach has been well documented for a wide range of intercalation systems.2R Problems with the interpretation can (27) Thompson, A. H.J . Electmehem. Soe. 1@7S,126,603. (28) h a n d , M. In Materials for Advoneed Batteries;Murphy, D. W., Braadhead, J.. Steele, B. C. H.. Eds.: Plenum: New York, 1980: p 145.
arise from nonequilibrium effects particularly when the free energy differences between adjacent phases are very small. In the present experiments, a small limiting current was used (10 p A ) and uncertainties due to nonequilibrium effects were assessed by measurements made on both oxidation and reduction cycles. All of the features seen in reduction are also observed on oxidation. The effect of using a finite limiting current shows up only as a relative offset in V and x of the reduction and oxidation data. Galvanostatic Measurements. Galvanostatic measurements were made a t a constant current of 50 p A (100 pA/cm2) between the voltage limits of -0.4 and +0.6V versus HgRIgO. Each time the temperature was changed the sample was reduced to a voltage cutoff of -0.5 V in order to obtain a stoichiometry as close as possible to LazNiO4.0. Measurements were made in the following temperature sequence: 273,283,295, and 268
K. The galvanostatic data at 273 K are shown in Figure 3 for one complete oxidation and reduction cycle. The data show that the oxidation-reduction reaction is reversible with respect to the change in the composition and there is no indication of 0 2 evolution on oxidation. The two-phase region observed at room temperatureL6 in the composition range 0.01 5 x 5 0.06 is also apparent in the data a t 273 K. The separation in voltage between the oxidation and reduction cycle a t x = 0.03is 0.12V. A new plateau region indicating two phase behavior is observed a t 273 K on oxidation between x = 0.08 and 0.12. The same feature is observed in the reduction cycle but is offset in composition. At room temperature a single phase is observed in this composition range. The polarization measured by the difference in voltage between the oxidation and reduction cycles is higher in this region (AV= 0.16a t x = 0.11) than in the range 0.01 5 x 5 0.06. The transition region between the two two-phase regions is complex and different between oxidation and reduction cycles. Clearly the reaction is far from equilibrium a t these current densities (100pA/cm2).
Intercalation of Oxygen in La.&i04+,
Chem. Mater., Vol. 6, No. 11, 1994 2081 0.4
0.3
t
II
0.2
. 2" E
--
0
v v)
0
>
-0.2 \
-0.4
x in LazNi04+x Figure 4. Galvanostatic reduction and oxidation curves (50 PA) for LazNiO4 a t 283 K.
The galvanostatic results at 283 K are essentially similar (Figure 41,though the data are significantly less polarized (AV = 0.086 at x = 0.03 and AV = 0.089 at x = 0.11). More than one change in slope is apparent between the low and high composition regions suggesting the possibility of more than one intermediate phase. When the temperature is lowered to 268 K, the polarization increases significantly and the galvanostatic data are no longer reversible in that the oxygen removed on reduction cannot quantitatively be reintercalated on oxidation. When the temperature is lowered from 295 to 268 K, the maximum composition that can be reached on oxidation decreases. The overpotential for 0 2 evolution also decreased as the experiments proceeded and consequently, the compositions measured at the upper cutoff voltage of 0.6 V are overestimates of the amount of intercalated oxygen. Values of x at 0.4 V (x = 0.148 a t 295 K, 0.137 at 283 K, 0.127 a t 273 K, and 0.116 at 268 K), well below oxygen evolution, are not sensitive to this effect and show the progression with temperature in the maximum composition that can be obtained on oxidation. The oxidation curves for the data at the four temperatures are shown in Figure 5. The differences in the maximum oxygen content that can be obtained at the different temperatures are probably due to the slower kinetics at lower temperatures. It is also possible that the phase boundary a t the high oxygen contents is strongly temperature dependent. Potential Step Experiments. Potential step experiments were carried out at 273 and 283 K on the same sample used for the galvanostatic experiments. These results will be compared with the data taken at 295 K on a separate electrode from the same batch of starting materials. The experiments used 5 mV steps in the voltage with a lower current cutoff of 10 PA. The ambient temperature data are shown in Figure 6 for reduction (top) and oxidation (bottom). The reduction data were reported previously.16 The derivative of the composition with respect to voltage (dxldv) is also
'
I
0
-Oa3r -0.51...'...'...'...'. .'-.'...I 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
283K
0.03
0.06
0.09
0.12
0.15
x in La,NiO,+x
Figure 5. Galvanostatic oxidation data for LazNiO4 a t 268, 273, 283, and 295 K. dxIdV 0
1
0.5
1.5
2
0.12
0.09 X
0.06
0.03 n
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Volts (vs (HglHgO) dx/dV
0 0.15
1
I
2
3
I
0.12
0.09 X
0.06
0.03 oxidation
0
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Volts (vs Hg/HgO)
Figure 6. Variation of voltage with composition (x) for reduction and oxidation of LazNiOd+, at 295 K obtained from potential step data. The reduction and oxidation data are shown in the top and bottom panels, respectively. In each case, peaks in the derivative (dxldv) correspond to the centers of two phase regions. shown. The derivative plot is useful in locating intermediate phases. Ideally, maxima in dxldV correspond to the midpoints in composition of two-phase regions. In practice, because of kinetic effects associated with diffusion and nucleation, the curves are broadened and
2082 Chem. Mater., Vol. 6, No. 11, 1994
Yazdi et al.
Table 1. Summary of the Voltage and Composition Data Obtained from the Potential Step Measurements at 273 and 283 K
oxidation 273 K x at (drldV),,
V x
0.045 -0.160
at (dVldx),,
V 283 K x at (&/dV),,
V x at (dVldx),, V
0.008 -0.186
0.082 -0.110 0.064 -0.124 0.072 -0.171 0.067 -0.131
0.106 0.123 0.000 f0.050 0.081 0.112 -0.051 -0.037 0.092 0.116 -0.036 0.037 0.083 0.110 -0.076 0.016
reduction 0.018 -0.210 0.05 -0.206
shifted in position between the oxidation and reduction cycles. The ambient temperature data are consistent with the phase diagram reported by Rice and B ~ t t r e y A . ~single ~ phase exists in the composition range from x = 0.06 to the upper composition limit that can be accessed before oxygen evolution occurs (x = 0.14-0.15). The singlephase region is characterized by a nearly constant value for dxldV. Below x = 0.06, a two phase region is indicated by a nearly constant voltage and a maximum in dx/dV. The compositions at the derivative peak maxima are 0.025 and 0.047 for the reduction and oxidation data, respectively. The average value is 0.036, slightly higher than the expected composition midpoint from the published equilibrium phase diagram. Exact correspondence of the average midpoint and the equilibrium midpoint requires that polarization effects are equal in each direction. In general, the pathway will contribute some difference to the kinetics and an exact average is not expected. A single phase with space group P4dncm and x = 0.03 reported by Rice and Buttrey21 is not observed in the present experiments using polycrystalline samples. The P4dncm phase has been observed in electrochemical experiments with a single crystal of LazNi04 and shown to exist only over ~ ~ formation of this a narrow range of c o m p o s i t i ~ n .The phase may play a role in the reaction kinetics in the two phase region below x = 0.06, even though it is not directly observed in data for polycrystalline electrodes. Data were collected at 273 K for two complete cycles of oxidation and reduction after an initial reduction t o fix the starting composition at x = 0. The data for the two complete cycles of oxidation and reduction were nearly identical. Data for the second oxidation and third reduction are shown together in Figure 7. The third reduction cycle data together with the derivative dxldV are shown in Figure 8 on an expanded scale. The oxidation and reduction data are shown in Figure 9 in the same format for comparison with the 295 K results in Figure 6. In addition t o the maximum in the derivative corresponding to the two-phase region below x = 0.06, three peaks in dxIdV are observed in the composition region that is single phase at 295 K (Figure 8). The maxima are observed in both oxidation and reduction but are more pronounced in the reduction data. Values of x and V corresponding t o the maxima on oxidation and reduction, and the average values are given in Table 1. In a similar way, maxima in the derivative dVldx can be used to locate the composition centers of the single phases between the two-phase regions. These data are (29) DiCarlo,J. F.; Yazdi, I.; Bhavaraju, S.; Jacobson, A. J.; Buttrey,
D. J., to be published.
0.070 -0.170 0.065 -0.187 0.081 -0.140 0.067 -0.176
average
0.094 0.115 0.032 -0.060 -0.015 -0.185 0.093 0.111 -0.124 -0.046 0.105 0.127 0.029 -0.047 0.012 -0.196 0.091 0.110 -0.105 -0.030
z . 2
0.076 -0.140 0.065 -0.156 0.077 -0.156 0.067 -0.154
0.100 0.030 0.087 -0.088 0.099 -0.042 0.087 -0.091
0.119 0.018 0.112 -0.005 0.122 0.025 0.11 -0.007
0.1
oxidation
*
in v
in
-
-0.10
3
,
0
>
f-
-Oe3
c-
I
0 0.02 0.04 0.06 0.08
x
reduction
0.1 0.12 0.14
in La,NiO,+x
Figure 7. Variation of voltage with composition for the second oxidation and the third reduction of LazNi04+, at 273 K obtained from potential step data.
0.5
l
r
3
8 . r” h
0.3
6 v
a x
= 2 c
0.1
-* m
0
> -0.1 1 -0.3
0
-0.5 0 0.02 0.04 0.06 0.08
0.1 0.12 0.14
x in La,NiO,+x
Figure 8. Variation of voltage with composition for the third reduction cycle reduction of L a ~ N i 0 4 + a t~273 K obtained from
potential step data. Peaks in the derivative (dr/dV)correspond t o the centers of two-phase regions.
not shown graphically, but values of x and V corresponding to the dV/dx maxima are given in Table I. The data show that the single-phase region at 295 K
Intercalation of Oxygen in LafliO*+,
Chem. Mater., Vol. 6,No. 11, 1994 2083
dxldV 0.14
0
,
1
dxldV
3
2
4
5
0
1
3
1
4
0.15
0.12
0.09 X
0.06
0.03
I
0 -0.4
-0.3
0
-0.1
-0.2
0.1
0.2
-0.4
F 0 ” l -0.3
-0.2
dx/dV 0
1
0.1
0.2
dx/dV 3
2
0
-0.1
Volts (vs Hg/HgO)
Volts (vs HglHgO)
0
5
4
0.14
3
2
I
5
4
I
0.04
0.03 0.02
-
i
! \ ,
oxidation
oxidation
Y
0 -0.4
.0.3
-0.2
-0.1
0
0.1
0.2
Volts (vs Hg/HgO)
-0.4
.0.3
-0.2
-0.1
-5.96lO”O.I
0.2
Volts (vs HgiHgO)
Figure 9. Variation of voltage with composition ( x ) for reduction and oxidation of LazNiO*+, at 273 K obtained from potential step data. The reduction and oxidation data are shown in the top and bottom panels, respectively. In each case, peaks in the derivative (dx/dV) correspond to the centers of two phase regions.
Figure 10. Variation of voltage with composition (x) for reduction and oxidation of LazNiOl+, at 283 K obtained from potential step data. The reduction and oxidation data are shown in the top and bottom panels, respectively. In each case peaks in the derivative (dx/dV) correspond to the centers of two phase regions.
separates into three two-phase regions with midpoint compositions corresponding to x = 0.08, 0.10, and 0.12. The resolution in these experiments is insufficient t o more precisely define the phase boundaries. Opencircuit voltage measurements at specific compositions are required to complete the phase diagram. The present data combined with the additional information provided by structural measurements enables the selection of a minimum set of compositions to further define the phase diagram. It should be noted, however, that the equilibration times needed to obtain open-circuit voltages are expected to be long. The present “quasiequilibrium” experiments typically took 11days for one complete cycle of reduction and oxidation. Similar potential step data were also measured at 283 K and are shown in Figure 10. Values of x and V corresponding to the maxima on oxidation and reduction and the average values are given in Table 1. Qualitatively the data at 283 and 273 K are similar (Figures 9 and 10) though the derivative peaks are not quite so sharp in the 283 K data. A more quantitative comparison can be made using the results in Table 1. The average compositions at which the maxima in both dxl dV and dVldx occur are closely similar indicating that the phase diagram is little altered in cooling from 283 to 273 K.
indicate that phase separation occurs in the composition range 0.06 I x I 0.14. The phase separation is indicated by the appearance of a plateau in galvanostatic data when the fully reduced composition is oxidized. The plateau is clearly apparent at 283 K and becomes even more pronounced at lower temperatures. When the temperature is lowered, the cell resistance increases and the intercalation rates decrease. A combination of the two effects limits the lowest temperature at which useful data can be obtained to 273 K. Below this temperature composition changes are not reversible between oxidation and reduction cycles. Potential step experiments provide more detail on the phase separation behavior for 0.06 Ix I 0.14. The kinetics are slow relative to the phase separation observed in the composition range 0 Ix I 0.06. The slower kinetics are indicated by the composition differences between corresponding peaks observed on oxidation and reduction cycles and are indicative of a complex structural reorganization. The data at 283 and 273 K are essentially the same. Maxima in the derivatives dxl dV and dVldx have been used to obtain compositions corresponding to the centers of the two-phase and the single-phase regions, respectively. The average values obtained for the oxidation and reduction cycles provide a good approximation to the equilibrium phase diagram assuming that the magnitude of the polarizations due to the reaction kinetics are the same on the oxidation and reduction cycles.
Discussion Electrochemical intercalation and deintercalation experiments on LaoNiOl below ambient temperature
2084 Chem. Mater., Vol. 6, No. 11, I994
o'2 0.1
Yazdi et al.
r-----
1
0.02
0.04
0.06
0.08
0.1
0.12
0.14
x in La,NiO,+, Figure 11. Schematic phase diagram for LaZNiOlh at 213 K, the shaded areas indicate two-phase regions.
The derivative plots show in Figures 9 and 10 show well-defined though broad maxima that are shifted in composition between the oxidation and reduction cycles. The number of peaks is the same for both cycles and the correspondence between the peaks in the oxidation and reduction data is clear. The width of the derivative peaks suggests that further resolution might be possible in even slower experiments. On the basis of the data in Table 1, a schematic phase diagram can be constructed. The exact phase boundaries are uncertain and have been estimated from the widths of the maxima in dddV. The results obtained by analysis of the 273 and 283 K data are almost identical. The schematic phase diagram is shown in Figure 11. The single-phase region observed a t ambient temperature separates a t 273 K into three two-phase regions with compositionscentered a t x = 0.076,0.100, and 0.119. Single phases with compositions centered a t x = 0.065, 0.087, and 0.112 separate the two-phase regions. We have shown previously that the electrochemically determined phase diagram is closely similar to the ambient temperature phase diagram determined using diffraction studies of crystals with different compositions.21 The different oxygen stoichiometries were prepared by annealing the crystals in different oxygen fugacities a t high temperatures. The electrochemical results below ambient temperature also can be compared with the recent neutron diffraction studies on single crystals of LazNiOa+, with oxygen stoichiometries where x = 0.105, 0.085, 0.074, 0.06, and 0.058. Very detailed diffraction studies of the phases present and of the kinetics of phase separation as a function of temperature have been reported.25 The data were interpreted in terms of the formation of staged compositions where layers containing interstitial oxygen atoms regularly alternate with specific numbers of empty
layers. The staging is apparently similar to that observed in graphite and the layered dichacogenides. A stange n compound has one filled layer regularly ordered with n - 1 empty layers. The phase diagram determined by the neutron diffraction studies for the LazNi04h system below ambient temperature can be compared with the electrochemical data. For x 5 0.06 the electrochemical data and the structural measurements show that the phase diagram is unaltered on cooling. The electrochemical data on polycrystalline electrode do not show the P4dncm phase though it has been observed in recent single-crystal electrochemical measurement^.^^ For x 2 0.06, the neutron diffraction data indicate phase separation into two phases: a stage 3 compound with 0.065 5 x 5 0.70; and a stage 2 compound with 0.10 5 x 5 0.11. These two phases correspond to the two single-phase regions determined from the electrochemical data with compositions centered at 0.065 and 0.11. The agreement is very close though not exact due to uncertainties in determining the boundaries in each experiment. In the electrochemical data, uncertainties in phase boundaries arise from the broadening of the data due to non-equilibrium effects. The phase limits determined structurally are limited by the number of measured compositions. Taking these factors into account, the agreement is very good. The electrochemical data also indicate the formation of a distinct phase with a composition centered at x = 0.087, not identified in the neutron diffraction study. The structural phase diagram proposed by Tranquada et al.25suggests that this composition should lie in a two-phase region, but compositions close to this value were not examined in the neutron diffraction experiments (the nearest were 0.085 and 0.105). Some evidence for the existence of a distinct phase at x = 0.09 has been obtained by synchrotron X-ray d i f f r a ~ t i o n . ~ ~ The electrochemical data indicate that further structural studies x = 0.087 would be of value to determine the nature of this phase. Initial studies at ambient temperature show that more detailed information can be obtained by using single-crystal electrodes. Electrochemical experiments currently in progress on single crystals at lower temperatures are expected to provide more information on the precise location of the phase boundaries. Electrochemicalintercalation provides very high composition resolution but only over a limited range of temperature. In contrast, the structural studies provide great detail over a wide range of temperature but a t specific compositions. The two techniques provide complementary information and together can be used to map out the detailed phase equilibria in oxide systems like LafliO4 that have remarkably high oxygen diffusivities.
Acknowledgment. We thank D. J. Buttrey for helpful discussions. The work was supported in part by the Robert A. Welch foundation (Grant No. 1207). (30) Buttrey, D.J.: Rice, D.E.,to be published
